Field Detail

• m_vModels

  protected java.util.Vector<CAAMModel> m_vModels

  Model vector.

• m_vImagePyr

  protected java.util.Vector<ModelSimpleImage> m_vImagePyr

  Image VOIs vector.

Constructor Detail

• CAAMModelMS

  public CAAMModelMS()

  Default multi-scale constructor.

Method Detail

• dispose

  public void dispose()

  dispose memory

  Overrides:

  dispose in class CAAMModel

• BuildAllLevels

  public void BuildAllLevels​(int nLevels,
                             java.lang.String inDir,
                             java.lang.String acf)

  Wrapper for automate model generation.

  Parameters:

  nLevels - number level in multi-scale

  inDir - input dir

  acf - aam configuration file

• BuildAllLevels

  public void BuildAllLevels​(int nLevels,
                             java.lang.String inDir,
                             java.lang.String acf,
                             int modelReduction,
                             int excludeShape)

  Driver method for model generation. This method automates the model generation as much as possible by using the various class methods for all the sequences in the task of producing a model.

  Parameters:

  nLevels - The number of levels the multi-scale AAM should be in. Default 3.

  inDir - Input directory where annotations (.asf) resides.

  acf - Filename of an AAM configuration file. If omitted defaults are used.

  modelReduction - Model reduction multiplier. Default off == 1. This sets the size of the lowest level in the model pyramid, i.e. level 0.

  excludeShape - Excludes one shape number 'excludeShape' from the input directory. Default -1, i.e. no shapes are removed. Used to perform leave-one-out testing.

• BuildAllLevels

  public void BuildAllLevels​(int nLevels,
                             java.util.Vector<java.lang.String> asfFiles,
                             java.lang.String acf,
                             int modelReduction,
                             int excludeShape)

  Driver method for model generation. This method automates the model generation as much as possible by using the various class methods for all the sequences in the task of producing a model.

  Parameters:

  nLevels - The number of levels the multi-scale AAM should be in. Default 3.

  asfFiles - Vector of asf filenames.

  acf - Filename of an AAM configuration file. If omitted defaults are used.

  modelReduction - Model reduction multiplier. Default off == 1. This sets the size of the lowest level in the model pyramid, i.e. level 0.

  excludeShape - Excludes one shape number 'excludeShape' from the input directory. Default -1, i.e. no shapes are removed. Used to perform leave-one-out testing.

• WriteModel

  public boolean WriteModel​(java.lang.String filename,
                            boolean txt_only)

  Writes the complete multi-scale AAMModel to disk as a set of .txt and an .amf files.

  Overrides:

  WriteModel in class CAAMModel

  Parameters:

  filename - Output filename without any extension. Multi-scale prefixes will be added.

  txt_only - If true binary model data is not written.

  Returns:

  true on success, false on file errors.

• ReadModel

  public boolean ReadModel​(java.lang.String basename)

  Reads the complete AAMModel from disk.

  Overrides:

  ReadModel in class CAAMModel

  Parameters:

  filename - Input filename without any extension and scale prefixes. E.g. if the files on disk are 'model_msl.txt' & 'model_msl.amf' -> filename = 'model'

  Returns:

  true on success, false on file errors.

• OptimizeModel

  public CAAMOptRes OptimizeModel​(ModelSimpleImage image,
                                  CAAMShape s,
                                  CDVector c,
                                  int maxIterations,
                                  java.util.Vector<CAAMOptState> pOptStates,
                                  boolean disableDamping)

  Performs AAM optimization of a shape containing initial pose using a model pyramid.

  Overrides:

  OptimizeModel in class CAAMModel

  Parameters:

  image - The image to search in (in size corresponding to level zero).

  s - The initial shape (also containing the inital pose, thus; not a normalized shape). Actually only the pose of 's' is used (to align the reference shape as the initial shape). NOTE: The optimal shape is returned in 's' after execution. NOTE: 's' is defined at level 0, i.e. shapes should have the same size as when calling CAAMModel::OptimizeModel() directly.

  c - The optimal model parameters at level zero. Unlike CAAMModel::OptimizeModel() the content of this vector is not used, since the optimisation is started the smallest (i.e. top) level of the pyramid. NOTE: The optimal model parameters are returned in 'c' after execution.

  maxIterations - The maximum iterations allowed at each level.

  pOptStates - Optional parameter convergence info (at level zero) can be returned in. See CAAMOptState.

  disableDamping - Disables the damping steps (default false).

  Returns:

  The results of the optimization in the form of a 'CAAMOptRes' instance.

• BuildPyr

  public void BuildPyr​(ModelSimpleImage image)

  Builds an image pyramid (if its not cached beforehand). The pyramid is stored in m_VImagePyr.

  Parameters:

  image - Input image that should be convertet to a pyramid. This image is assumed to live to the *complete* usage of the pyramid, since level 0 is a pointer to this image.

• Model

  public final CAAMModel Model​(int i)

  Get the model at level i.

  Parameters:

  i - level

  Returns:

  model

• NLevels

  public final int NLevels()

  Get the number of levels.

  Returns:

  levels number

• GetSmallest

  public final CAAMModel GetSmallest()

  Get the smallest model

  Returns:

  model